Main Analysis Dataset
Summary Statistics for the Main Analysis Dataset
## user_id status_id created_at screen_name
## Length:1063 Length:1063 Length:1063 Length:1063
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## text source display_text_width NUMYearsTwitterUser
## Length:1063 Length:1063 Min. : 6.0 Min. : 0.000
## Class :character Class :character 1st Qu.: 75.0 1st Qu.: 2.000
## Mode :character Mode :character Median :140.0 Median : 6.000
## Mean :150.4 Mean : 5.537
## 3rd Qu.:232.0 3rd Qu.: 9.000
## Max. :301.0 Max. :13.000
##
## BINSocialInfluencer CATUserIndividualOrGroup BINTextContainsEmojis
## Min. :0.0000 Length:1063 Min. :0.000
## 1st Qu.:0.0000 Class :character 1st Qu.:0.000
## Median :0.0000 Mode :character Median :0.000
## Mean :0.3104 Mean :0.175
## 3rd Qu.:1.0000 3rd Qu.:0.000
## Max. :1.0000 Max. :1.000
##
## TXTEmojiFound CATCopyRightMaterial CATLonelinessType CATTemporalBounding
## Length:1063 Length:1063 Length:1063 Length:1063
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## CATSocialContext BINPhysicalContext BINRomanticContext BINSomaticContext
## Length:1063 Min. :0.0000 Min. :0.0000 Min. :0.0000
## Class :character 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000
## Mode :character Median :0.0000 Median :0.0000 Median :0.0000
## Mean :0.1505 Mean :0.1505 Mean :0.4911
## 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:1.0000
## Max. :1.0000 Max. :1.0000 Max. :1.0000
##
## CATDerivedAgeGroup BINCommInteraction BINCOVID
## Length:1063 Min. :0.0000 Min. :0.0000
## Class :character 1st Qu.:0.0000 1st Qu.:0.0000
## Mode :character Median :1.0000 Median :0.0000
## Mean :0.5616 Mean :0.3076
## 3rd Qu.:1.0000 3rd Qu.:1.0000
## Max. :1.0000 Max. :1.0000
##
## BINAdvocatingAwarenessForLoneliness BINProjectToOther retweet_count
## Min. :0.0000 Min. :0.0000 Min. : 0.0000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.: 0.0000
## Median :0.0000 Median :0.0000 Median : 0.0000
## Mean :0.2832 Mean :0.1656 Mean : 0.9981
## 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.: 0.0000
## Max. :1.0000 Max. :1.0000 Max. :449.0000
##
## hashtags place_full_name location followers_count
## Length:1063 Length:1063 Length:1063 Min. : 0
## Class :character Class :character Class :character 1st Qu.: 123
## Mode :character Mode :character Mode :character Median : 384
## Mean : 2612
## 3rd Qu.: 1448
## Max. :200594
## NA's :1
## friends_count listed_count statuses_count favourites_count
## Min. : 0.0 Min. : 0.00 Min. : 1 Min. : 0
## 1st Qu.: 197.2 1st Qu.: 1.00 1st Qu.: 1170 1st Qu.: 1093
## Median : 520.0 Median : 4.00 Median : 4604 Median : 5567
## Mean : 1131.6 Mean : 40.56 Mean : 15455 Mean : 16332
## 3rd Qu.: 1112.8 3rd Qu.: 22.00 3rd Qu.: 16197 3rd Qu.: 18841
## Max. :135394.0 Max. :3511.00 Max. :510189 Max. :331637
## NA's :1 NA's :1 NA's :1 NA's :1
## urls_expanded_url account_created_at status_url
## Length:1063 Length:1063 Length:1063
## Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
##
##
## CATUserDerivedLocation CATTweetCreatedDay TIMTweetCreatedTime long
## Length:1063 Length:1063 Min. : 0.000 Min. :144.4
## Class :character Class :character 1st Qu.: 4.000 1st Qu.:145.0
## Mode :character Mode :character Median : 8.000 Median :145.0
## Mean : 9.292 Mean :144.9
## 3rd Qu.:13.000 3rd Qu.:145.0
## Max. :23.000 Max. :145.1
##
## lat CATAnalysis_1 CATAnalysis_2 CATAnalysis_3
## Min. :-38.15 Length:1063 Length:1063 Length:1063
## 1st Qu.:-37.82 Class :character Class :character Class :character
## Median :-37.82 Mode :character Mode :character Mode :character
## Mean :-37.83
## 3rd Qu.:-37.82
## Max. :-37.74
##
## CATAnalysis_4
## Length:1063
## Class :character
## Mode :character
##
##
##
##
How Many Twitter Users are Represented, by Location?
Pinpointing exact locations of where a Tweet geographically originated on can be technically tricky. What we can generally say from this view is, most of the Tweets have originated from “close” to the Geelong region, within at least 100km radius.
| Tweet Location | User Type | Total Users | Total Tweets | Average Years Twitter User |
|---|---|---|---|---|
| EastMetro | Individual | 6 | 7 | 4.57 |
| GeelongBellarine | Individual | 17 | 20 | 5.05 |
| InnerMelbourne | Business | 18 | 22 | 5.91 |
| InnerMelbourne | Group | 21 | 78 | 4.91 |
| InnerMelbourne | Individual | 756 | 905 | 5.62 |
| NorthMetro | Business | 1 | 3 | 4 |
| NorthMetro | Individual | 5 | 6 | 7 |
| SouthEastMetro | Individual | 5 | 6 | 6.17 |
| Unknown | Individual | 10 | 10 | 3.1 |
| WestMetro | Individual | 6 | 6 | 5 |
Top Twitter Users by Number of Tweets About Loneliness
Using the “screen_name” variable we can tally up the distinct Twitter users represented in the dataset.
It’s encouraging to see “@friendsfor_good”, “@EndLonelinessAU” and “@Dr_FaithG” on this list. These are Twitter users who advocate for raising awareness about the loneliness subject.
| User | Total Tweets |
|---|---|
| @friendsfor_good | 37 |
| @shaunrowland11 | 11 |
| @DigitalMehmet | 8 |
| @EndLonelinessAU | 5 |
| @barbaraneves | 4 |
| @Dr_FaithG | 4 |
| @fakenotears | 4 |
| @myronmy9 | 4 |
| @Nashtopia | 4 |
| @pixelsmixel | 4 |
Timeline of Tweets
Acknowledging the timeline for capturing data, a 6 week period, is quite a short amount of time. In 2020, with the COVID-19 pandemic and it especially severely impacting Victoria, it is curious to note that around the first week of July and first week of August there are spikes in the numbers of Tweets about loneliness. These time points also coincide with Victorian State Government announcements about COVID-19 public safety and lock-down measure announcements. Melbourne was placed on stage-4 restrictions, while regional Victoria was placed on stage-3.
Tweets at Times Throughout the Day
This density plot shows us that the highest volume of Tweets were sent out during the 4AM to 10AM time frame, with a lowest volume at 5PM with a small increase at 10PM.
This is interesting to observe, however without full context may not be as informative.
Tweets Throughout the Week
This bar plot shows that Wednesday’s and Thursday’s during the week were the most popular days for Tweets about loneliness. It is curious to observe that the weekend days were the least popular days for Tweeting about loneliness.
Locations and Geographical Origins of Tweets
Most Twitter users turn off their location in their privacy settings but those that don’t add valuable location information to their Tweets.
By using “place_full_name” supplemented with “location”, it was possible to manually allocate geographical metropolitan and regional areas to generally categorise Tweet locations.
In the table below, we can observe that “InnerMelbourne” represents the majority of Tweet locations, followed by “GeelongBellarine”. It is possible that “GeelongBellarine” is under-represented here as users may opt to display the nearest capital city (Melbourne) as their location rather than precisely where they geographically are. It is also equally possible that mobile phones, laptops and personal computer devices may be using the nearest cell tower or internet service provider network server to give a general geographical position at the time of Tweet creation.
In the map below, the dark green circle represents a 50km radius around the Geelong CBD, the geographical parameter used in the Twitter API data extraction. The instances of number of Tweets per general Tweet location are overlayed. The groups “Unknown” and “GeelongBellarine” have been combined in this view.
#Tweet Locations
#Data Setup
#Definition Location patterns used in grepl of CATUserDerivedLocation in wrk_100_main_analysis.
loc_unknown <- c("#purpose: shelter my soul", "any pronouns", "she/her")
loc_eastmetro <- c("Blackburn", "Box Hill", "Abbotsford", "Hawthorn", "Ringwood", "Surrey Hills")
loc_northmetro <- c("Brunswick", "Coburg", "Epping", "Northcote")
loc_westmetro <- c("3018", "Footscray", "Melton", "Point Cook","Werribee")
loc_southeastmetro <- c("Clayton", "Elwood", "Murrumbeena", "Frankston","Mornington Peninsula")
loc_geelongbellarine <- c("Geelong", "Lara", "Barwon Heads", "Drysdale","Clifton Springs","Portarlington","Ocean Grove")
loc_regionalvic <- c("Colac")
loc_surfcoast <- c("Angelsea")
loc_innermelb <- c("mel", "Melbourne", "MELBOURNE", "Albert Park", "Carlton", "Docklands","Prahan","Somewhere in Melbourne","South Yarra")
DerivedLocations <- c("Unknown", "EastMetro","NorthMetro", "WestMetro","SouthEastMetro", "GeelongBellarine", "RegionalVic", "SurfCoast", "InnerMelbourne" )
lat <- c(-38.148743,-37.813680, -37.741317, -37.859042, -37.905374, -38.148743, -38.342728, -38.408793, -37.817007)
long <- c(144.365739,145.089686, 144.966709, 144.796066, 145.125195, 144.365739, 143.584951, 144.162215, 144.951542)
geocoordslocation <- data.frame(DerivedLocations,long,lat ) # This set is left joined to wrk_100_main_analysis by DerivedLocations = CATUserDerivedLocation
wrk_100_main_analysis %>%
filter(!is.na(CATUserDerivedLocation)) %>%
count(CATUserDerivedLocation, sort = TRUE) #Geographical Plot of user locations:
# Setting up summary dataset for geographical plotting:
# Combine the Unknown location with Geelong, else when we plot it, it will overlap.
wrk.geoplot <- wrk_100_main_analysis %>%
mutate(CATUserDerivedLocation = ifelse(CATUserDerivedLocation == "Unknown", "GeelongBellarine",CATUserDerivedLocation)) %>%
group_by(long, lat, CATUserDerivedLocation) %>%
summarise(NUMUsers = n_distinct(user_id),
NUMTweets = n() ) # Create groups to plot
userlocations.df <- split(wrk.geoplot, wrk.geoplot$CATUserDerivedLocation)
pal <- colorFactor(c("purple", "navy", "turquoise", "orange", "gold", "green"),
domain = c("InnerMelbourne", "GeelongBellarine", "NorthMetro", "EastMetro", "SouthEastMetro", "WestMetro"))
#Define the leaflet object
#leaflet(wrk.geoplot) %>% addTiles() %>% addMarkers(
# clusterOptions = markerClusterOptions()
#)
l <- leaflet() %>% addTiles()
names(userlocations.df) %>%
purrr::walk( function(df) {
l <<- l %>%
addProviderTiles(provider = "CartoDB.Positron") %>%
addLabelOnlyMarkers(data = userlocations.df[[df]], lng = ~long, lat = ~lat,
label = ~as.character(NUMTweets) ,
labelOptions = labelOptions(noHide = T, direction = 'center', textOnly = T, textsize = "20px",
style = list(
"color" = "black",
"font-family" = "arial",
"font-style" = "bold"))) %>%
addCircles(lng = 144.3657, lat = -38.14874, radius = 50000, col = "green", fill=FALSE) %>%
addCircleMarkers(data = userlocations.df[[df]],
radius = 45,
color = ~pal(CATUserDerivedLocation),
lng = ~long, lat = ~lat,
stroke = FALSE,
fillOpacity = 0.5,
label = ~as.character(CATUserDerivedLocation),
popup = ~as.character(NUMTweets),
group = df)
})
# Create UI control layer for the map plots
l %>%
addLayersControl(
overlayGroups = names(userlocations.df),
options = layersControlOptions(collapsed = FALSE)
)
How Much Text Content Per Tweet?
We can inspect “display_text_width” to observe how many text characters are used per Tweet.
The maximum character limit for Twitter is 280 characters. The median for this dataset is 140 characters per Tweet, approximately one sentence worth, per Tweet.
What Devices are Used to Communicate Tweets?
We can inspect “source” to observe the devices which are used to Tweet from.
This list reveals there are many device, linkage and platform options to “Tweet” from.
At the top of this list it is evident that users are using an iPhone, Laptop/PC or Android smartphone to use Twitter.
| Tweet Sources | Total Users | Total Tweets |
|---|---|---|
| Twitter for iPhone | 370 | 440 |
| Twitter Web App | 219 | 284 |
| Twitter for Android | 181 | 220 |
| Twitter for iPad | 23 | 24 |
| Hootsuite Inc. | 13 | 17 |
| TweetDeck | 11 | 14 |
| 10 | 12 | |
| Buffer | 5 | 9 |
| 3 | 5 | |
| IFTTT | 3 | 4 |
| dlvr.it | 2 | 3 |
| Sprout Social | 2 | 3 |
| Tweetbot for iS | 3 | 3 |
| Fenix 2 | 1 | 2 |
| Goodreads | 2 | 2 |
| How You Really Feel | 1 | 2 |
| Microsoft Power Platform | 2 | 2 |
| Tweetbot for Mac | 1 | 2 |
| Echobox | 1 | 1 |
| Falcon Social Media Management | 1 | 1 |
| Flamingo for Android | 1 | 1 |
| 1 | 1 | |
| Grabyo | 1 | 1 |
| HubSpot | 1 | 1 |
| Lightful | 1 | 1 |
| Missinglettr | 1 | 1 |
| Ripl App | 1 | 1 |
| Scoop.it | 1 | 1 |
| SEMrush Social Media Tool | 1 | 1 |
| SocialBee.io v2 | 1 | 1 |
| Streamlabs Twitter | 1 | 1 |
| Twitter for Mac | 1 | 1 |
| WordPress.com | 1 | 1 |
The Most Re-Tweeted Tweets
Even though the Twitter API search query was configured not to extract re-Tweets, we have access to ReTweet counts in the data.
The Twitter API helpfully returns a “retweet_count” variable whose values can easily be sorted. Here we sort all the Tweets in descending order by the size of the “retweet_count”.
| Created | User | ReTweet Count | Text |
|---|---|---|---|
| 2020-07-31T13:23:00Z | trashcanprince | 449 |
even an evil manifestation gets lonely sometimes #neiboltreddie #reddie https://t.co/UDamH44Oak |
| 2020-07-30T09:37:00Z | luciemorrismarr | 51 | This is my neighbour. He’s 85, a widower and lives alone with his dog. He finds life lonely but today he said through his window he feels relief at his life choices. “I’m not in aged care, I’m alive,” he says. #COVID19Vic https://t.co/HMnj2ycNak |
| 2020-07-08T00:30:00Z | ReadingsBooks | 48 | Okay! Because we know it can be loneliness inducing thinking about the next six weeks, here are six things we can all do at home to feel part of the bookish community and deal with that feeling of isolation … |
| 2020-07-28T21:44:00Z | MazinB_ | 42 | Always pick being alone rather than settle for someone that makes you feel lonely |
| 2020-07-30T05:12:00Z | barbaraneves | 24 | When we wrote this piece on experiences of lonely older Australians we couldn’t anticipate how things would be now: so much worse! Our participants are reporting suicidal ideation and complete despair, because they’re not only lonely, they feel disposable & blamed. Ageism kills! https://t.co/0UoWG3zpo5 |
| 2020-07-31T01:31:00Z | ItsSpoopsB | 23 | Scoops feeling a bit lonely https://t.co/kpovubxElJ |
| 2020-07-07T10:01:00Z | philipdalidakis | 19 | Please take care out there & look out for each other. Another lock down will be challenging to many. Financial pressures, loneliness/isolation, family violence. But you’re not alone. Call Lifeline 131144 or Beyond Blue 1300 22 4636 24 hours a day. We will only beat this together! |
| 2020-08-02T09:53:00Z | DrEricLevi | 13 |
Dear Melbourne, You can be in a crowd and feel lonely. You can be alone and not feel lonely. Connect well during this season. We can be physically distant but be socially and emotionally connected. |
| 2020-07-14T07:28:00Z | MartinFoleyMP | 13 | Victorians struggling with loneliness will now be able to receive support from @RedCrossAU and community organisations thanks to the $6 million Community Activation and Social Isolation initiative, funded through our $59.4 million mental health and wellbeing package #springst https://t.co/wHh9wswdtN |
| 2020-07-16T01:44:00Z | NormanHermant | 12 |
At #agedcareRC devastating evidence from 91 yr old Beryl Hawkins via telephone. She lives on her own in a Housing Commission flat in #Sydney. She experienced loneliness & depression, ‘sitting for hrs not being able to talk to anyone.’ Also terrible #dental problems. -Thread- https://t.co/EVpLvSRPNk |
It is possible to extract a visual screen-shot of the re-Tweeted Tweets, using Tweet_screenshot() from the Tweetrmd package. Just provide the “screen_name” and “status_id”.
Re-Tweeted Tweets About Loneliness
The following set of Tweet screenshots are from the top 10 re-Tweeted Tweets about loneliness.
Tweets with the Most Likes
To find the most liked Tweet we can sort the Tweets by the “favorite_count” variable in descending order and print the rows with the highest counts.
By observing the number of re-Tweets and likes, we can begin to assess the social “reach” that Tweets about loneliness may have on the social media community and network of associated people.
| Created | User | Favourites Count | Text |
|---|---|---|---|
| 2020-08-01T21:49:00Z | DamienEvans7 | 331637 | @mrsnb16 @PatsKarvelas The loneliness of the long distance runner - Iron Maiden |
| 2020-08-04T00:59:00Z | ewster | 321567 |
Gaaaaah they be buzzing about again, can hear the bickering, and the lonely guy keeps zooming up and down Elizabeth St. He needs a date. Tough going under lockdown. |
| 2020-07-07T05:52:00Z | RV_27 | 313584 | @Origsmartassam You look a bit lonely Sam |
| 2020-07-30T05:17:00Z | Wolfie_Rankin | 276877 |
@LugubriousLarry Thankyou kindly. I’m grateful for those who are online with me, but in real life I’m quite lonely. I miss my family terribly. This old house used to be a very busy place, people would randomly drop in, the phone was always going. And now most of the time there’s silence. |
| 2020-07-17T10:53:00Z | gerster_kaylene | 276638 | @Gary_Hardgrave @EndlessEcho121 Sadly these stories of lonely grief will last longer than virus…lots of pent up anger will cause many problems in the future.. |
| 2020-07-07T07:08:00Z | gerster_kaylene | 275671 | @barrelracernt @bouta_nt Bored and lonely… |
| 2020-08-05T09:55:00Z | stofsk | 254978 | @damoj Maybe I’m just sad and lonely but that wouldn’t be a turn off for me |
| 2020-07-09T08:33:00Z | andrewfx_51 | 211098 | @Asher_Wolf @Sunsplashsun Serious pros and cons to the idea. A lot of classicists/fundamentalists do it because you can devote greater time to classical languages. For someone who was “gifted” the more important skills I learned were not academic, although it was very lonely also |
| 2020-07-16T03:00:00Z | missannaklein | 197815 | @BMiyakee @lenacarti Lost in an image, in a dream But there’s no one there to wake her up - the world is spinning But tell me what happens when it stops? They go And they say She’s so lucky, she’s a star But she cry, cry, cries in her lonely heart, thinking If there’s nothing missing in my life… |
| 2020-07-31T09:41:00Z | mareefeb | 172230 | @greyham65 @chelsea_hetho no one dies before their time the majority of the deaths have been to people already dying of other complaints if covid really was the cause of death who can tell?…it is the lonely isolated deaths that is the problem & anyone who thinks that is okay is just cruel |
Most Liked Tweets About Loneliness
The following set of Tweet screenshots are from the top 10 most liked Tweets about loneliness.
Which Twitter Users are Mentioned the Most?
Here we tokenise the text of each Tweet and use str_detect() from the stringr package to filter out words that start with an @ .
In this list we observe the Premier of the State of Victoria, Dan Andrews, Friends for Good, ABC Melbourne and The Age newsmedia agency.
| Mentions User Name | Mentions Count |
|---|---|
| @DanielAndrewsMP | 14 |
| @rpatulny | 6 |
| @dariusdevas | 5 |
| @theage | 5 |
| @friendsforgood | 4 |
| @JessTu2 | 4 |
| @MarleeBower | 4 |
| @abcmelbourne | 3 |
| @AcadSocSciences | 3 |
| @bairdjulia | 3 |
Screenshots Tweets Containing Users Who are Frequently Mentioned
The Tweet screen-shot below is an example of a response to a new book which had been released in Victoria. A new author named Jessie Tu (see top mentions: @JessTu2) has written and published a book titled “A Lonely Girl is a Dangerous Thing”. This book not only contains “lonely” in the title, a key word in this analysis, but it is actually a book which includes the topic of loneliness and evokes a strong response from within the book reading community. There are numerous Tweets in the analysis data which reference this book title and the topics it covers.
Summary of Research Variables
Linking back to Section 6b where features were created to enhance the analysis datasets, the variables which were inspired by the works of Kivran-Swaine, F, et al (2014) and Ruiz, C, et al (2017) will be used to form categorical groups which can be used in the subsequent emoji and text analysis.
The primary analysis grouping that this analysis will use for subsequent sections is comprised of a concatenation of “Temporal Bounding” (CATTemporalBounding) and “Loneliness Type” (CATLonelinessType), this gives us 6 categorical groups, a fairly good number to commence analysis with. In contrast, when this categorical grouping is combined with the Loneliness Context variables (BINPhysicalContext, BINRomanticContext, BINSomaticContext), Communication Interaction (BINCommInteraction), COVID-19 flag (BINCOVID), Advocating Awareness for Loneliness (BINAdvocatingAwarenessForLoneliness) and Projecting Loneliness (BINProjectToOther), the cardinality of the categories is 57, which is too many for most statistical summaries to visualise neatly.
The tables below describe the combined analysis groups and show the cardinality (unique combinations).
| Temporal Bounding and Loneliness Type | Total Users | Total Tweets |
|---|---|---|
| ambiguous_social_loneliness | 484 | 631 |
| enduring_social_loneliness | 228 | 251 |
| transient_social_loneliness | 64 | 68 |
| enduring_individual_loneliness | 48 | 52 |
| ambiguous_individual_loneliness | 47 | 49 |
| transient_individual_loneliness | 12 | 12 |
| Loneliness Context | Total Users | Total Tweets |
|---|---|---|
| seekinginteraction | 221 | 272 |
| NoContext | 168 | 228 |
| _somatic__seekinginteraction_ | 135 | 148 |
| somatic | 127 | 138 |
| _physical__somatic_seekinginteraction | 64 | 70 |
| _romantic__somatic_seekinginteraction | 53 | 56 |
| _romantic__somatic_ | 34 | 37 |
| _physical__somatic_ | 31 | 32 |
| _physical__romantic__somatic__seekinginteraction_ | 28 | 29 |
| _romantic__seekinginteraction_ | 11 | 13 |
| _physical__romantic_somatic | 11 | 12 |
| romantic | 11 | 11 |
| _physical__seekinginteraction_ | 8 | 8 |
| physical | 7 | 7 |
| _physical__romantic_ | 1 | 1 |
| _physical__romantic_seekinginteraction | 1 | 1 |
| Loneliness Alternative Context | Total Users | Total Tweets |
|---|---|---|
| NoAltContext | 397 | 442 |
| covid | 143 | 157 |
| lonelinessadvocate | 87 | 148 |
| _covid__lonelinessadvocate_ | 105 | 140 |
| projectingloneliness | 125 | 135 |
| _covid__projectingloneliness_ | 26 | 28 |
| _lonelinessadvocate__projectingloneliness_ | 10 | 11 |
| _covid__lonelinessadvocate_projectingloneliness | 2 | 2 |
| Full Descriptor | Total Users | Total Tweets |
|---|---|---|
| ambiguous_social_loneliness_seekinginteraction_ | 206 | 254 |
| ambiguous_social_loneliness | 158 | 217 |
| enduring_social_loneliness_somatic_ | 63 | 69 |
| enduring_social_loneliness_somatic_seekinginteraction | 63 | 66 |
| ambiguous_social_loneliness_somatic_seekinginteraction | 47 | 54 |
| enduring_social_loneliness_physical__somatic__seekinginteraction_ | 46 | 50 |
| ambiguous_social_loneliness_somatic_ | 43 | 45 |
| transient_social_loneliness_somatic_seekinginteraction | 21 | 23 |
| enduring_social_loneliness_physical_somatic | 21 | 22 |
| transient_social_loneliness_somatic_ | 20 | 22 |
| enduring_individual_loneliness_romantic__somatic__seekinginteraction_ | 18 | 19 |
| ambiguous_individual_loneliness_romantic__somatic__seekinginteraction_ | 15 | 16 |
| ambiguous_social_loneliness_physical__somatic__seekinginteraction_ | 14 | 16 |
| enduring_individual_loneliness_physical__romantic__somatic_seekinginteraction | 10 | 11 |
| enduring_individual_loneliness_romantic_somatic | 10 | 11 |
| enduring_social_loneliness_romantic__somatic__seekinginteraction_ | 11 | 11 |
| enduring_social_loneliness_physical__romantic__somatic_seekinginteraction | 10 | 10 |
| ambiguous_social_loneliness_romantic_somatic | 8 | 9 |
| ambiguous_individual_loneliness_romantic_somatic | 8 | 8 |
| ambiguous_social_loneliness_physical_seekinginteraction | 8 | 8 |
| enduring_social_loneliness_seekinginteraction_ | 8 | 8 |
| ambiguous_individual_loneliness_romantic_seekinginteraction | 7 | 7 |
| ambiguous_social_loneliness_physical_ | 7 | 7 |
| ambiguous_social_loneliness_physical_somatic | 7 | 7 |
| enduring_social_loneliness_physical__romantic__somatic_ | 6 | 7 |
| ambiguous_individual_loneliness_romantic_ | 6 | 6 |
| enduring_social_loneliness | 5 | 6 |
| transient_individual_loneliness_romantic_somatic | 6 | 6 |
| ambiguous_individual_loneliness_seekinginteraction_ | 4 | 5 |
| ambiguous_social_loneliness_romantic_seekinginteraction | 3 | 5 |
| transient_social_loneliness_romantic__somatic__seekinginteraction_ | 5 | 5 |
| ambiguous_individual_loneliness_somatic_seekinginteraction | 4 | 4 |
| ambiguous_social_loneliness_physical__romantic__somatic_seekinginteraction | 4 | 4 |
| transient_social_loneliness | 4 | 4 |
| transient_social_loneliness_seekinginteraction_ | 4 | 4 |
| ambiguous_social_loneliness_romantic__somatic__seekinginteraction_ | 2 | 3 |
| enduring_individual_loneliness_physical__romantic__somatic_ | 3 | 3 |
| transient_social_loneliness_physical__romantic__somatic_seekinginteraction | 3 | 3 |
| transient_social_loneliness_physical_somatic | 3 | 3 |
| transient_social_loneliness_physical__somatic__seekinginteraction_ | 3 | 3 |
| ambiguous_individual_loneliness_physical__romantic__somatic_ | 2 | 2 |
| enduring_individual_loneliness_romantic_ | 2 | 2 |
| enduring_social_loneliness_romantic_somatic | 2 | 2 |
| transient_individual_loneliness_romantic_ | 2 | 2 |
| transient_individual_loneliness_romantic__somatic__seekinginteraction_ | 2 | 2 |
| ambiguous_individual_loneliness | 1 | 1 |
| ambiguous_social_loneliness_physical__romantic__seekinginteraction_ | 1 | 1 |
| ambiguous_social_loneliness_romantic_ | 1 | 1 |
| enduring_individual_loneliness_physical_romantic | 1 | 1 |
| enduring_individual_loneliness_physical__somatic__seekinginteraction_ | 1 | 1 |
| enduring_individual_loneliness_romantic_seekinginteraction | 1 | 1 |
| enduring_individual_loneliness_seekinginteraction_ | 1 | 1 |
| enduring_individual_loneliness_somatic_ | 1 | 1 |
| enduring_individual_loneliness_somatic_seekinginteraction | 1 | 1 |
| transient_individual_loneliness_physical__romantic__somatic_seekinginteraction | 1 | 1 |
| transient_individual_loneliness_somatic_ | 1 | 1 |
| transient_social_loneliness_romantic_somatic | 1 | 1 |
Summary of Created Features
The tables below show the summary statistics for the features created in the main analysis dataset. This analysis does not cover these variables in further detail.
| Derived Age Group | Total Users | Total Tweets |
|---|---|---|
| Adult | 814 | 1010 |
| Youth_Student | 29 | 32 |
| Children | 11 | 21 |
| Social Influence | Total Users | Total Tweets |
|---|---|---|
| Low Influence | 590 | 733 |
| High Influence > 1000 followers | 255 | 330 |
| Copyright Material Type Referenced | Total Users | Total Tweets |
|---|---|---|
| None | 682 | 807 |
| Other | 86 | 148 |
| Quote | 51 | 57 |
| SongLyric | 34 | 37 |
| FilmTVTitle | 14 | 14 |
| Social Context | Total Users | Total Tweets |
|---|---|---|
| Online | 558 | 700 |
| Offline | 315 | 363 |